Salmonella serovars along two beef chains in Ethiopia

Introduction: Salmonella has been reported from foods and the food production environment, with outbreaks occurring in the human population worldwide. Methodology: A survey on Salmonella in two beef production lines (a beef abattoir line and a processing line) in Addis Ababa, Ethiopia was conducted, with a total of 668 various samples randomly collected from animal-related materials, the environment, and a beef product (mortadella). Results: Overall, a 12.9% prevalence (26.3% from the abattoir line, 5.3% from the processing plant line) was observed. The prevalence in the abattoir line environment (36.6%) was higher than that in animal-related samples (14.7%); the reverse was true for the processing plant line. Out of 86 isolates, 10 serovars were identified, and 8 remained unidentified. The predominant serotypes were S. Saintpaul (32.5%), S. Muenchen (19.8%), and S. Larochelle (12.8%). S. Kastrup and S. London were isolated for the first time in Ethiopia. Conclusions: Data indicate open ports of entry for Salmonella, with possible transfer along the line. Further investigations from farm to fork are recommended in order to identify these positions of entry.


Introduction
In Ethiopia, consumption of raw meat is traditional, which carries the risk of foodborne infections and intoxications [1].Several studies on Salmonella prevalence in Ethiopia have been published (e.g., among cattle, slaughterhouse personnel, the environment, and minced beef) [1][2][3].
Previous studies in Ethiopia did not follow structured sampling plans along beef production chains.This survey intended to identify possible sources for transfer of Salmonella serovars along two beef chains in Ethiopia.

Ethical considerations
This project was approved and funded by the Ethiopian Engineering Capacity Building Program (ECBP) offices of Ethiopia.

Abattoir and processing lines
A cross-sectional study was carried out along two lines: a cattle abattoir line and a beef processing plant line.In the abattoir line, multi-purpose cattle stocks purchased from extensive or semi-intensive management systems in different parts of the country, either tracked or trucked, are slaughtered at Addis Ababa Abattoir Enterprise (AAAE) [17].The AAAE has a capacity of up to 1,200 cattle in 8 hours with a staff of about 700 persons.After slaughter, carcasses are delivered to city butcheries, immediately or after a short cooling interval.Butcheries are mostly small open-stall shops, handling the meat at 20°C-27°C, which is the ambient temperature in Addis Ababa City.
The processing plant line is located at Bishoftu town, 47 km east of Addis Ababa.It receives raw beef from three abattoirs, (from AAAE in Addis Ababa, from Adama Municipal Abattoirs in Adama [located 90 km east of Addis Ababa] and from Bishoftu Municipal Abattoirs in Bishoftu).In this small-scale plant (8 to 10 working persons), beef is processed in a working area without intersections either immediately or is kept in a refrigerator until processing.The product (beef mortadella) goes to private supermarkets in Addis Ababa City.Here, products are kept in the refrigerator with other products of animal origin.Slicing is done using one slicing machine for all products during supply to the consumers.

Sampling
Samples were taken from December 2011 to April 2012 over 18 sampling occasions: 5 times from the abattoir and butchery, and 13 times from the beef processing plant line, 8 times from the processing plant, and 5 times from supermarkets.
In the abattoir, samples were taken from the operation environment, directly from the animal/product, and raw beef from city butchery locations.In the processing plant line, samples were taken from the environment, from animal products, and from supermarkets in Addis Ababa.
From both lines, 668 samples from a total of 35 sampling locations were taken (Table 1).For swabs, a 50 cm 2 area was swabbed with sterilized gauze moistened with normal saline solution.Water (20 mL) was filled directly from the tap into sterile calibrated glass bottles.Tissue and product samples were taken aseptically and placed in sterile stomacher bags.Samples were immediately transported to Microbiology Laboratory, Akililu Lemma Institute of Pathobiology, Addis Ababa University (ALIPB-AAU), Ethiopia on the day of sampling using an ice box at +4°C.

Sample preparation
Each sample was aseptically taken.For pre-enrichment, buffered peptone water (Merck, Darmstadt, Germany) was used.The first 1:10 dilution was homogenized with a Stomacher 400 (Seward Laboratory, London, UK) and incubated at 37°C for 18-20 hours to be used as pre-enrichment [18,19].

Data analysis
Data were entered in to Microsoft Excel 2007 (Microsoft Corp., Redmond, USA) and analyzed using Excel, State 11, and SPSS version 20 (IBM Corp., Armonk, USA).Percentage and mid-prevalence exact 95% confidence intervals (CI) were used to demonstrate prevalence differences between and among the sampling occasions and types of samples.

Serovars
In total, 86 Salmonella strains were obtained and serotyped (Table 4).Of the 10 different serovars identified, 3 of them were found only in the abattoir line and 4 only in the processing plant line.A total of 3 serovars (S.Saintpaul, S. London, S. Muenchen) along with unidentified ones were isolated from both lines.Predominant serotypes were S. Saintpaul (32.5%), S. Muenchen (19.8%), and S. Larochelle (12.8%).Abattoir line S. Saintpaul was the predominant serotype (11.4%; 95% CI: 7.8-15.9),being present in all sampling locations with the exception of hooks (Table 4).S. Saintpaul was followed by S. Muenchen (5.9%) and S. Larochelle (4.6%).S. Dublin was observed only in room samples, animal feces, and raw meat at the abattoir and in the butcheries.
At the fifth sampling occasion, only two serotypes (S.Muenchen and S. Larochelle) were observed (Table 5).S. Dublin and S. London were observed only during the first sampling occasion.At the fourth sampling, the frequency of S. Muenchen and S. Larochelle was higher in environmental samples than in animal-related material and samples from the butcheries (0 isolates; Table 5).
With respect to sampling location, hand of personnel, aprons, and knives were frequently positive with S. Saintpaul, S. Muenchen, and S. Larochelle.

Processing plant line
S. London was obtained 5 times at the second sampling occasion, S. Eastbourne was observed at the fifth, seventh, and eighth sampling.All others were obtained only infrequently (Table 6).
S. Eastbourne was observed in samples from the environment and from animal-related material, while S. Muenchen was detected in raw meat and in the end product (supermarket).S. London was observed mostly in raw meat.

Prevalence
The overall prevalence of 12.9% in this study was lower than that reported by Molla et al. [10], with 23.6% from food animals in Ethiopia, and similar to a study reporting 7.1% positive samples out of 323 cattle in Debre Zeit [8].

Abattoir line
The abattoir line produced a 26.6% prevalence, which was higher than the 10.9% reported by Sibhat et al. [3] in Ethiopia and also higher than the 7.2% prevalence reported by Teklu and Negussie [6] in a sheep and goat abattoir line at Modjo, Ethiopia, which was erected more recently and which possesses a clear and transparent technical line.
The occurrence of Salmonella in all sampling locations and occasions along this line may be due to the continuous influx of animals that contaminate abattoir and the equipment (floors, personnel) during processing.
Positive results from the lymph nodes indicate the infection status of the animals.The 8.8% result in the present study was similar with 8% reports of Sibhat et al. [3], but higher than the 4.2% found in slaughter  The abattoir environment was more frequently positive (36.6%) than were animal-related samples (14.7%).Incoming strains may establish themselves as a permanent in-house flora under poor cleaning and disinfecting conditions [7].

Environmental samples
Positive environmental samples ranged between 30.7% in knives and 60% in refrigerators.All results indicate heavy cross-contamination, which is true for people as well as for the surfaces of tools and equipment.
Among water samples, 8.3% were positive, similar to the results of Teklu and Negussie [6] in water used at Modjo abattoir (7.1%).Samples taken from the trucks were more frequently positive (45.5%), with rates as high as those of butcheries.

Samples from beef and butcheries
The prevalence in beef at the abattoir level (11.8%) was similar with the 9.8% rate reported by Nyeleti et al. [1], higher than the 2.8% and 3.1% rates [8], 2% [3] from carcass swabs at a beef abattoir, and lower than 42.8% (n = 236) reported from Senegal [21].It was similar to the rates of 11.9% and 9.8% found in the diaphragm and abdominal muscles, respectively [1].
The 32.4% positive results in raw beef at the butcheries were similar to findings at the abattoir, lower than the 87.4% rate reported by Stevens et al. [21] from retail beef in Senegal and the 60% rate found among samples from a South African slaughterhouse [22].
The number of positive samples at the butcheries and in animal-related material was high as well, indicating possible transfer from the abattoir into the butcheries and from there into the human habitat.

Processing plant line
Only a few environmental locations were positive along the processing plant line.Of these, the 5.2% prevalence observed from personal hand swabs at the processing plant was similar to the 7% rate reported by Sibhat et al. [3] and the 10.6% reported by Teklu and Negussie [6] from hand samples.The 17.7% prevalence obtained from working tables was lower than the 96.4% at permanent markets and 70% at districts sales places on wood and cardboard [21].Cutting plates, floors, as well as cutters and stuffers were positive as well.

Meat and mortadella
Starting with a 10.5% Salmonella prevalence in the processing plant, in 119 mortadella samples, only 1 sample was positive.
In comparison, higher positive numbers were obtained from raw products, e.g., 7.9% in minced beef [1], 14.4% in minced beef, 14.1% in mutton, and 16.4% in pork from a supermarket in Addis Ababa [2].Our data indicate the different kind of commodities exposed to Salmonella contamination risk.The application of heat treatment destroys Salmonella and lowers the risk of product contamination.
The 1.2% S. Typhimurium proportion was lower than the 20% reported by Alemayehu et al. [8].Investigations in children from Addis Ababa and Jimma [13] yielded S. Typhimurium in 0.8% and 0.3% of cases, respectively, with an overall prevalence of 0.7% in hospitals.
The presence of S. Dublin was slightly higher than the 2.4% reported by Ejeta et al. [2] but lower than the 48% reported by Alemayehu et al. [8].
The 2.3% proportion of S. Anatum found in this study is similar to the 2.6% reported by Molla et al. [9] from camels, but lower than the 9.1% reported by Ejeta et al. [2].S. Anatum was the most reported serotype (62.1%) in the study of Sibhat et al. [3].This was also the case in a study from an abattoir in Algiers, Algeria [23], where S. Anatum was the predominant serotype among Salmonella isolates.
S. Saintpaul was the predominant serotype isolated in this investigation.The present 32.5% S. Saintpaul percentage was similar with the 38.8% reported by Molla et al. [9] in camels, but lower than the 2.3% reported by Ejeta et al. [2].Comparing both abattoir and processing line, the prevalence of S. Saintpaul was higher in the abattoir line than at the processing plant line (0.25%; 95% CI: 0.01-1.13)(p < 0.05).
The 19.8% proportion of S. Muenchen was higher than the 8.6% reported by Molla et al. [9] from camels and higher than the 0.7% from pigs reported by Aragaw et al. [14] in Ethiopia.
The prevalence of S. Concord was low (0.3%).Beyene et al. [13] reported an overall 4.2% prevalence with 5.2% at Addis Ababa and 2.3% at Jimma Hospitals as a major pathogen in children with diarrhea in Ethiopia.
The present 0.5% positive sample for S. Eastbourne was lower than the 15/278 in caecal contents, 21/278 in MLN, and 3/277 in carcass swabs reported by Aragaw et al. [14] from pigs at abattoir in Ethiopia.
Sibhat et al. [3] obtained S. Eastbourne from cattle hides, MLN, and from a carcass surface.

Conclusions
Salmonella has been reported from foods and the food production environment, with outbreaks occurring in the human population worldwide.
The structured survey presented in this study was aimed at Salmonella serotypes' detection in two beef production lines (a beef abattoir line and a processing line) in Addis Ababa, Ethiopia.The study results indicated the presence of this agent in animal-related materials, in the abattoir line environment, and in a heat-treated beef product (mortadella).The application of heat treatment to the screened products, which is able to destroy Salmonella during the steam cooking operation, lowers the risk of contamination.
Finally, isolation of S. Kastrup and S. London for the first time in Ethiopia also suggests the possible presence of diversified Salmonella serotypes.Hence, national based Salmonella surveys in food and food production and processing lines are recommended.

Table 1 .
Sampling locations, sample types, and number enrolled from both lines.

Table 3 .
Salmonella isolates by sampling location and type of samples (processing line).

Table 4 .
Salmonella serovars obtained from abattoir and processing plant samples.

Table 5 .
Salmonella serovars by sampling location and occasion (abattoir line).

Table 6 .
Distribution of Salmonella serovars in positive sampling locations and occasions at processing plants.